[jsumrall]

Amazon Q. Claude Code is great (the best imho, what everything else measures against right now), and Amazon Q seems almost as good and for the first week I've been using it I'm still on the free tier.

The flat pricing of Claude Code seems tempting, but it's probably still cheaper for me to go with usage pricing. I feel like loading my Anthropic account with the minimum of $5 each time would last me 2-3 days depending on usage. Some days it wouldn't last even a day.

I'll probably give Open AI's Codex a try soon, and also circle back to Aider after not using it for a few months.

I don't know if I misundersand something with Cursor or Copilot. It seems so much easier to use Claude Code than Cursor, as Claude Code has many more tools for figuring things out. Cursor also required me to add files to the context, which I thought it should 'figure out' on its own.

[wordofx]

> I don't know if I misundersand something with Cursor or Copilot. It seems so much easier to use Claude Code than Cursor, as Claude Code has many more tools for figuring things out. Cursor also required me to add files to the context, which I thought it should 'figure out' on its own.

Cursor can find files on its own. But if you point it in the right direction it has far better results than Claude code.

[jatins]

this is the first time I am seeing someone says good things about Amazon Q

Do they publish any benchmark sheet on how it compares against others?

[retinaros]

It is currently at top3 in swe bench verified.

It went through multiple stages of upgrades and I would say at this stage it is better than copilot. Fundamentally it is as good as cursor or windsurf but lacks some features and cannot match their speed of release. If you re on aws tho its a compelling offering.

[quintes]

I remember asking Amazon Q something and it wouldn’t reply cuz of security policy or something. It was as far as I can remember a legit question around Iam policy which I was trying to configure. I figured it out back in Google search.